Sinusoidal endothelial cell-derived FGF23 modulates hematopoiesis in mice Free

Fibroblast growth factor 23 (FGF23), a phosphaturic hormone produced by osteocytes, has been found to negatively regulate erythropoiesis ^{[PMID:24509850]}, while FGF23 C-terminal cleaved peptides lower hepcidin and raise circulating iron levels in mice ^{[PMID:37053547]}. Previously, using Fgf23 reporter mice, we defined bone marrow (BM) sinusoidal endothelial cells (SEC) as a novel site of Fgf23 upregulation in anemic mouse models, and we found that both large-volume phlebotomy and erythropoietin (EPO) injection increased Fgf23 reporter expression in BM-SEC ^{[PMID:37417950]}. Here we investigate the phenotypic consequences of genetic Fgf23 disruption in SEC in normal mice and in anemic mouse models with perturbed iron metabolism.

By mining single-cell RNA-seq data from BM stromal cell populations of inbred C57BL/6 mice ^{[PMID:31130381]}, we identified both Stabilin-2 (Stab2) and Fgf23 mRNA in the SEC cell cluster. Therefore, to disrupt Fgf23 expression in SEC, we bred Fgf23 conditional knockout mice (Fgf23^fl/fl)^{[PMID:26792657]} to mice that express Cre recombinase under the Stab2 promoter, which confers Cre activity in SEC of liver, spleen, lymph node, and BM ^{[PMID:28218627]}. Using the Stab2-iCreF2 founder line, we found that the ability of EPO injection to induce Fgf23 in total BM mRNA and to raise total plasma FGF23 was abolished in mice with conditional Fgf23 loss in SEC. To further define effects of Fgf23 loss in SEC, we utilized the Stab2-iCreF3 founder line, which confers Cre activity in SEC but not in hematopoietic cells or their progeny ^{[PMID:34845225]}. We confirmed Stab2-iCreF3 activity in adult BM vasculature by confocal microscopy, using a reporter allele that expresses tdTomato fluorescence following Cre-mediated recombination. Fgf23^fl/flStab2-iCreF3 mice (referred to hereafter as Fgf23^SEC/SEC) were healthy and born in expected Mendelian ratios. In 9-week-old Fgf23^SEC/SEC females, blood hemoglobin levels (Hgb) trended higher compared to sex-matched Fgf23^fl/fl littermate controls (2-tailed unpaired t test P=.05; n≥8), while body and spleen weights were similar. However, 5 days after 500μl phlebotomy, Fgf23^SEC/SEC females showed more severe reductions in blood red cell count and Hgb (P<.05) relative to baseline levels and a trend towards increased spleen-to-body weight ratio (P=.09) when compared to Fgf23^fl/fl females.

We next examined the effects of Fgf23 disruption in SEC in anemic mouse models that show plasma FGF23 elevation and Fgf23-reporter allele upregulation in BM SEC. In the Hbb^th3/+ model of non-transfusion dependent β-thalassemia, pups with Fgf23 loss in SEC (Hbb^th3/+Fgf23^SEC/SEC) were underrepresented by nearly 50% relative to Hbb^th3/+Fgf23^fl/fllittermates at genotyping on postnatal day 7. In a preliminary cohort (8-9 weeks old), Hbb^th3/+Fgf23^SEC/SEC females showed greater body weight (P<.05), lower mean corpuscular hemoglobin concentration (P<.05), and a trend towards lower plasma total FGF23 (mean 1295 vs 4276 pg/ml; P=.10) when compared to Hbb^th3/+Fgf23^fl/fl females (Mann-Whitney test; n=3-6). Bone histomorphometry revealed trends towards greater osteoblasts per total area, greater osteoid thickness, and greater osteoid volume per bone volume (P=.14) in Hbb^th3/+Fgf23^SEC/SEC females compared to Hbb^th3/+Fgf23^fl/fl females. By contrast, in the Tmprss6^-/- model of iron refractory iron deficiency anemia, Fgf23 loss in SEC did not markedly impact pup survival. When compared to sex-matched Tmprss6^-/-Fgf23^fl/fl littermate controls at 8 weeks, Tmprss6^-/-Fgf23^SEC/SEC females showed a modest but significant (P<.05) rise in Hgb but similar body and spleen weights (unpaired t test; n=4-10).

In summary, our findings suggest that in healthy mice, SEC-derived FGF23 exerts relative minor effects on steady-state hematopoiesis but contributes significantly to hematological recovery after large volume blood loss, a period where hepcidin suppression is required to increase iron availability for stress erythropoiesis. Additionally, our preliminary analysis suggests that SEC-derived FGF23 may impair growth of Hbb^th3/+ mice, a model in which ineffective erythropoiesis promotes hepcidin suppression and systemic iron loading. However, SEC-derived FGF23 does not appear to impact growth of Tmprss6^-/- mice, a model in which a failure to downregulate hepcidin signaling causes iron-restricted erythropoiesis. Studies evaluating the impact of SEC-derived FGF23 to iron homeostasis and bone health in these anemic models are ongoing.